*** zerocoolback has joined #kata-dev | 00:00 | |
*** LinuxMe has quit IRC | 00:11 | |
*** zerocoolback has quit IRC | 00:19 | |
*** LinuxMe has joined #kata-dev | 00:19 | |
*** zerocoolback has joined #kata-dev | 00:19 | |
*** LinuxMe has quit IRC | 00:23 | |
*** LinuxMe has joined #kata-dev | 00:24 | |
*** zerocoolback has quit IRC | 00:26 | |
*** zerocoolback has joined #kata-dev | 00:27 | |
*** LinuxMe has quit IRC | 00:29 | |
*** zerocoolback has quit IRC | 00:30 | |
*** zerocoolback has joined #kata-dev | 00:32 | |
*** zerocoolback has joined #kata-dev | 00:32 | |
*** LinuxMe has joined #kata-dev | 00:36 | |
*** annabelleB has quit IRC | 00:39 | |
*** LinuxMe has quit IRC | 00:41 | |
*** zerocoolback has quit IRC | 01:05 | |
*** zerocoolback has joined #kata-dev | 01:05 | |
*** zerocoolback has quit IRC | 01:06 | |
*** zerocoolback has joined #kata-dev | 01:09 | |
*** zerocoolback has quit IRC | 01:13 | |
*** dlw has joined #kata-dev | 01:45 | |
*** zerocoolback has joined #kata-dev | 01:51 | |
*** mcastelino has quit IRC | 02:02 | |
*** zerocoolback has quit IRC | 02:03 | |
*** zerocoolback has joined #kata-dev | 02:27 | |
*** zerocoolback has quit IRC | 02:45 | |
*** zerocool_ has joined #kata-dev | 02:45 | |
*** zerocool_ has quit IRC | 03:43 | |
*** zerocoolback has joined #kata-dev | 03:46 | |
*** Jeffrey4l has quit IRC | 03:48 | |
*** zerocoolback has quit IRC | 03:54 | |
*** Jeffrey4l has joined #kata-dev | 03:55 | |
kata-dev-irc-bot | <bergwolf> @stefanha Does fedora act different? I did test initramfs with alpine and centos rootfs and they do not require udev to create `/dev/virtio-ports`, which thus I assumed is created by the kernel instead. | 04:31 |
---|---|---|
*** mcastelino has joined #kata-dev | 04:38 | |
*** sjas_ has joined #kata-dev | 04:50 | |
*** sjas has quit IRC | 04:53 | |
*** LinuxMe has joined #kata-dev | 04:53 | |
*** djinni has quit IRC | 05:04 | |
*** LinuxMe has quit IRC | 05:11 | |
*** djinni has joined #kata-dev | 05:15 | |
*** dlw1 has joined #kata-dev | 05:26 | |
*** dlw has quit IRC | 05:27 | |
*** dlw1 is now known as dlw | 05:27 | |
*** eernst has quit IRC | 05:56 | |
*** sameo_ has joined #kata-dev | 06:09 | |
*** jodh has joined #kata-dev | 06:10 | |
*** jodh has quit IRC | 06:10 | |
*** jodh has joined #kata-dev | 06:10 | |
*** sameo_ has quit IRC | 06:25 | |
kata-dev-irc-bot | <caihaomin> @bergwolf @james.o.hunt @eric.ernst https://github.com/kata-containers/runtime/pull/299 ready to merge~~ CI is green now! thx~ | 06:56 |
*** mcastelino has quit IRC | 07:04 | |
*** sameo has joined #kata-dev | 07:51 | |
*** zerocoolback has joined #kata-dev | 07:52 | |
*** zerocoolback has quit IRC | 07:56 | |
*** zerocoolback has joined #kata-dev | 07:56 | |
*** davidgiluk has joined #kata-dev | 07:59 | |
*** zerocoolback has quit IRC | 09:51 | |
*** zerocoolback has joined #kata-dev | 10:37 | |
*** dlw has quit IRC | 11:23 | |
*** devimc has joined #kata-dev | 11:58 | |
*** fuentess has joined #kata-dev | 12:36 | |
*** eernst has joined #kata-dev | 12:53 | |
stefanha | Any tips on debugging the Kata VM if QEMU doesn't launch or the guest powers down quickly? In those cases I don't get useful syslog output from the runtime. | 12:57 |
stefanha | Ideally I'd like to see the VM's console output and disable the "quiet" kernel parameter. | 12:57 |
stefanha | So far I've run Kata's QEMU command-line manually so I could debug things. It works manually now but is still silent when run under Kata, so I need to start debugging it within Kata. | 12:59 |
*** sameo has quit IRC | 13:01 | |
devimc | stefanha, https://gist.github.com/devimc/f8e3373d8d8e6b3b5b56441d04875f2a | 13:10 |
devimc | could be useful | 13:10 |
stefanha | devimc: Thanks! I was wondering what console.sock was for and never got any output from it :) | 13:11 |
devimc | stefanha, if you want to attach to console.sock you need to disable debugging in config file | 13:12 |
stefanha | devimc: Great, will try that! | 13:12 |
devimc | and you need a custom kata image | 13:12 |
devimc | take a look to osbuilder | 13:12 |
stefanha | devimc: Okay, I am already building my own. | 13:12 |
stefanha | devimc: Why is a custom image necessary? I remember the debug options being in the kata runtime configuration file, so how does the image come into play? | 13:13 |
devimc | stefanha, you need bash in the image to attach to console.sock and have a prompt | 13:14 |
devimc | also once you have the image | 13:14 |
devimc | you should modify the kata service | 13:14 |
devimc | https://gist.github.com/devimc/f8e3373d8d8e6b3b5b56441d04875f2a#to-debug-cc-agent-through-consolesock-add-next-line-to-cc-agenttarget-cc-2x-or-clear-containerstarget-cc-30 | 13:14 |
stefanha | devimc: I see, thanks! | 13:14 |
devimc | to include a service that will start bash | 13:14 |
devimc | once you are attached to console.sock you will be in the same level than the agent | 13:16 |
*** eernst has quit IRC | 13:17 | |
*** eernst has joined #kata-dev | 13:18 | |
devimc | stefanha, in kata the path to the target file is different | 13:19 |
devimc | /usr/lib/systemd/system/kata-containers.target | 13:19 |
devimc | stefanha, https://paste.fedoraproject.org/paste/5LSwt-W7Tf3sGLiYa2m6rQ | 13:19 |
devimc | and this is my debug-console.service | 13:20 |
devimc | https://paste.fedoraproject.org/paste/l~BALZFI5T102W-o-5tXag | 13:20 |
stefanha | devimc: Thanks. QEMU is terminating too quickly, I think the agent is refusing to start and the guest powers down. | 13:21 |
stefanha | I guess I'll modify the runtime to remove "nowait" from -chardev console.sock. I want it to wait until I have the console in place :) | 13:22 |
*** dklyle has quit IRC | 13:22 | |
stefanha | QEMU is disappearing too quickly and the console output is lost. | 13:22 |
*** dlw has joined #kata-dev | 13:30 | |
devimc | stefanha, what version of qemu are you using ? | 13:30 |
devimc | stefanha, also you need to update all components | 13:31 |
devimc | running docker or k8s? | 13:31 |
stefanha | devimc: Docker | 13:31 |
devimc | umm ok that's weird | 13:32 |
devimc | enable debugging and check the journal | 13:32 |
devimc | sudo journalctl -b -t kata-proxy | 13:32 |
devimc | sudo journalctl -b -t kata-runtime | 13:32 |
stefanha | devimc: I'm not using a stock Kata, I'm building from source and trying the Fedora rootfs. | 13:32 |
stefanha | And distro QEMU, so I'm not expecting things to work out-of-the-box. | 13:33 |
devimc | your host is RH ? | 13:33 |
stefanha | Was just wondering if there are smarter ways to debug than what I've been doing :) | 13:33 |
stefanha | devimc: yes | 13:33 |
devimc | ouch! | 13:33 |
devimc | I'm not sure if RH QEMU works with kata | 13:37 |
stefanha | devimc: Right | 13:39 |
devimc | if qemu version is > 2.7 it should works | 13:40 |
devimc | stefanha, you could build it from source and try | 13:41 |
devimc | https://github.com/kata-containers/qemu/tree/qemu-lite-2.11.0 | 13:41 |
stefanha | devimc: Yes, building QEMU from source should work. | 13:41 |
jodh | stefanha: the debug console info is now in the dev guide: https://github.com/kata-containers/documentation/blob/master/Developer-Guide.md#set-up-a-debug-console | 13:42 |
jodh | stefanha: I'm surprised you don't get any useful messages if you've enabled *full* debug? https://github.com/kata-containers/documentation/blob/master/Developer-Guide.md#enable-full-debug | 13:42 |
* stefanha checks for agent.log=debug | 13:43 | |
stefanha | I was missing agent.log=debug, thanks! | 13:43 |
*** diga has joined #kata-dev | 13:49 | |
*** annabelleB has joined #kata-dev | 13:51 | |
*** eernst has quit IRC | 13:52 | |
*** dlw has quit IRC | 14:01 | |
*** zerocoolback has quit IRC | 14:14 | |
*** sameo has joined #kata-dev | 15:13 | |
*** eernst has joined #kata-dev | 15:16 | |
*** annabelleB has quit IRC | 15:18 | |
*** zerocoolback has joined #kata-dev | 15:45 | |
*** mrcastel has quit IRC | 15:51 | |
*** mrcastel has joined #kata-dev | 15:51 | |
*** eernst has quit IRC | 15:57 | |
*** eernst has joined #kata-dev | 15:57 | |
*** dklyle has joined #kata-dev | 15:58 | |
*** devimc has quit IRC | 15:59 | |
*** eernst_ has joined #kata-dev | 16:01 | |
*** eernst has quit IRC | 16:02 | |
*** eernst has joined #kata-dev | 16:03 | |
*** eernst has quit IRC | 16:03 | |
*** eernst_ has quit IRC | 16:03 | |
*** eernst has joined #kata-dev | 16:03 | |
*** zerocoolback has quit IRC | 16:09 | |
*** eernst has quit IRC | 16:12 | |
*** eernst has joined #kata-dev | 16:13 | |
*** eernst_ has joined #kata-dev | 16:14 | |
*** eernst has quit IRC | 16:15 | |
*** eernst_ has quit IRC | 16:15 | |
*** eernst has joined #kata-dev | 16:15 | |
*** eernst has quit IRC | 16:19 | |
*** eernst has joined #kata-dev | 16:19 | |
*** eernst has quit IRC | 16:20 | |
*** zerocoolback has joined #kata-dev | 16:22 | |
*** eernst has joined #kata-dev | 16:22 | |
*** annabelleB has joined #kata-dev | 16:45 | |
*** mrcastel has quit IRC | 16:47 | |
*** mrcastel has joined #kata-dev | 16:49 | |
*** david-lyle has joined #kata-dev | 16:52 | |
*** devimc has joined #kata-dev | 16:54 | |
*** dklyle has quit IRC | 16:56 | |
*** david-lyle has quit IRC | 16:56 | |
*** dklyle has joined #kata-dev | 16:57 | |
*** jodh has quit IRC | 17:03 | |
*** zerocoolback has quit IRC | 17:07 | |
*** zerocoolback has joined #kata-dev | 17:07 | |
*** zerocoolback has quit IRC | 17:13 | |
*** zerocoolback has joined #kata-dev | 17:13 | |
*** zerocoolback has quit IRC | 17:13 | |
*** zerocoolback has joined #kata-dev | 17:14 | |
*** zerocoolback has quit IRC | 17:14 | |
*** zerocoolback has joined #kata-dev | 17:15 | |
*** zerocoolback has joined #kata-dev | 17:15 | |
*** zerocoolback has quit IRC | 17:16 | |
*** zerocoolback has joined #kata-dev | 17:16 | |
*** zerocoolback has quit IRC | 17:17 | |
*** zerocoolback has joined #kata-dev | 17:17 | |
*** zerocoolback has quit IRC | 17:17 | |
*** diga has quit IRC | 17:19 | |
*** david-lyle has joined #kata-dev | 17:55 | |
*** dklyle has quit IRC | 17:58 | |
*** mrcastel has quit IRC | 18:00 | |
*** dklyle_ has joined #kata-dev | 18:03 | |
*** david-lyle has quit IRC | 18:06 | |
*** mrcastel has joined #kata-dev | 18:21 | |
*** mrcastel has quit IRC | 18:23 | |
*** mrcastel has joined #kata-dev | 18:24 | |
*** dklyle_ has quit IRC | 18:35 | |
*** annabelleB has quit IRC | 19:01 | |
*** annabelleB has joined #kata-dev | 19:03 | |
*** annabelleB has quit IRC | 19:04 | |
*** dklyle has joined #kata-dev | 19:25 | |
*** LinuxMe has joined #kata-dev | 19:28 | |
*** LinuxMe has quit IRC | 19:42 | |
*** davidgiluk has quit IRC | 19:56 | |
*** devimc has quit IRC | 20:34 | |
*** annabelleB has joined #kata-dev | 20:42 | |
*** fuentess has quit IRC | 20:43 | |
*** fuentess has joined #kata-dev | 20:44 | |
*** fuentess has quit IRC | 21:34 | |
*** eernst has quit IRC | 21:57 | |
*** eernst has joined #kata-dev | 21:57 | |
*** eernst has quit IRC | 21:57 | |
*** eernst has joined #kata-dev | 21:58 | |
*** eernst has quit IRC | 21:58 | |
*** LinuxMe has joined #kata-dev | 22:03 | |
*** fuentess has joined #kata-dev | 22:10 | |
*** eernst has joined #kata-dev | 22:11 | |
*** LinuxMe has quit IRC | 22:13 | |
*** eernst has quit IRC | 22:14 | |
*** eernst has joined #kata-dev | 22:14 | |
*** eernst has quit IRC | 22:18 | |
*** eernst has joined #kata-dev | 22:18 | |
*** eernst has quit IRC | 22:23 | |
*** LinuxMe has joined #kata-dev | 22:26 | |
*** LinuxMe has quit IRC | 22:30 | |
*** fuentess has quit IRC | 22:49 | |
kata-dev-irc-bot | <tallclair> FYI - I've just added a section to my sandboxes doc analyzing the attack surfaces that would be exposed to sandboxed pods: https://docs.google.com/document/d/1QQ5u1RBDLXWvC8K3pscTtTRThsOeBSts_imYEoRyw8A/edit#heading=h.x07ordtme9n. | 23:00 |
kata-dev-irc-bot | <tallclair> Feedback on this would be very much appreciated! | 23:00 |
*** annabelleB has quit IRC | 23:14 | |
*** sameo has quit IRC | 23:22 | |
*** annabelleB has joined #kata-dev | 23:53 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!